I've built direct coupled 2a3 ampls like the Shinshido a couple of times. Sure it is easy, and sure the OPT's make a huge difference. Really good OPT's pretty much all cost money...
The true Loftin-white circuit (including the Daris version which uses a directly coupled cathode follower) is sounding pretty much better than any other direct coupled 2a3 I have put together before... I'm very happy with it right now 😀 😀 😀
Don't forget the tubes. I am using a bunch of OLD and cheap golden dragon chinese 2a3's that a friend gave me for free... I find that some of these old crappers need extra time to heat up, are pretty microphonic and even the heaters are noisy..
NOS are not much better mind you - fleabay is not a good source.. sellers these days pass off junk as NOS.. Decent quality current production would be wise for a final build.
The true Loftin-white circuit (including the Daris version which uses a directly coupled cathode follower) is sounding pretty much better than any other direct coupled 2a3 I have put together before... I'm very happy with it right now 😀 😀 😀
Don't forget the tubes. I am using a bunch of OLD and cheap golden dragon chinese 2a3's that a friend gave me for free... I find that some of these old crappers need extra time to heat up, are pretty microphonic and even the heaters are noisy..
NOS are not much better mind you - fleabay is not a good source.. sellers these days pass off junk as NOS.. Decent quality current production would be wise for a final build.
Good OPT are important and expensive BUT good power transformers can make an even bigger difference.
I mean, everybody looks for quality when designing and building an output transformer. Even an average Tango or Tamura unit offers decent audio quality.
Power transformers seldom receive adequate attention.
They are actually just as important as OPTs.
And should be built like OPTs...
I mean looking for low stray capacitance, etc.
A power transformer should share at least the same quality of the OPT.
I design my own transformers so I have plenty of possibility to make tests, and I have learned that it is better do not accept too many compromises.
Agreed.
L-W circuit has different current flows and souns better than Shishido's.
I also built a Shishido like dc coupled circuit in the beginning than modified it to have a different current flow.
That was so many years ago... no dc coupled amps in my system in the last decade
Of course cheap tubes are... cheap 🙂
but all parts are important: resistors, caps, chokes. All can make a huge difference.
I use cheap parts that I know well to breadboard and test. When it works well I then move on to some nicer components...
Not all expensive parts make a difference though. These ERO MKT 1813 polyester caps are very good -and they are not expensive. Same goes for the ASC X386s caps.
Even when bread-boarding I use good quality Mains and Chokes though... no reason to go cheap in the power supply...
If I was in the US I would be tempted to get some Elekta-Prints or similar nice quality iron. I have these terrible cheapest-ever tomiko's from a trip to japan a few years back.. got them just to test circuits with... they are ugly rotten sounding things.. but in this circuit sound pretty darn good... Even the golden dragons have settled down and are making beautiful music.
Not all expensive parts make a difference though. These ERO MKT 1813 polyester caps are very good -and they are not expensive. Same goes for the ASC X386s caps.
Even when bread-boarding I use good quality Mains and Chokes though... no reason to go cheap in the power supply...
If I was in the US I would be tempted to get some Elekta-Prints or similar nice quality iron. I have these terrible cheapest-ever tomiko's from a trip to japan a few years back.. got them just to test circuits with... they are ugly rotten sounding things.. but in this circuit sound pretty darn good... Even the golden dragons have settled down and are making beautiful music.
You live in Switzerland, don't you?
It's not far from Italy, I can have excellent OPTs wound here. I can source a pair for you if you want. Just send me a PM in that case.
On the original Loftin-white circuit from 1929, you might notice there are no cathode by-pass capacitors. A simple yet clever network is used to cancel PS noise.
It is possible to calculate when you won't need to use cathode by-pass. Darius called this the "triode amp characteristic factor". I will call it TAC for sake of simplicity. If this TAC factor is higher than 4, then you don't need to by-pass the cathode. Its really just a realtive number...
This TAC factor can be calculated with either of these ways:
Ra/((Rk+1/s)*mu) where s=transconductance
Ra/((mu*Rk)+ri) where ri= annode to cathode resistance
Basically, if you make sure that your anode resistor is big enough, then you don't need to bypass the cathode.
For High gain triodes you will however find that you need an anode resistor that is simply too large for your power supply. But Darius found a nice solution for this too.
In my amplifier I used Darius' method using 6SL7 to get a TAC that is an astonishing 40. 🙂
My6SL7/2A3 Loftin-White amp is dead silent and has truly superb bass response using only cheap 6 watt OPT's. Incredible imaging with no coupling caps or cathode by-pass used at all.
I should probably do some more measurements and maybe create a thread...
Ian
It is possible to calculate when you won't need to use cathode by-pass. Darius called this the "triode amp characteristic factor". I will call it TAC for sake of simplicity. If this TAC factor is higher than 4, then you don't need to by-pass the cathode. Its really just a realtive number...
This TAC factor can be calculated with either of these ways:
Ra/((Rk+1/s)*mu) where s=transconductance
Ra/((mu*Rk)+ri) where ri= annode to cathode resistance
Basically, if you make sure that your anode resistor is big enough, then you don't need to bypass the cathode.
For High gain triodes you will however find that you need an anode resistor that is simply too large for your power supply. But Darius found a nice solution for this too.
In my amplifier I used Darius' method using 6SL7 to get a TAC that is an astonishing 40. 🙂
My6SL7/2A3 Loftin-White amp is dead silent and has truly superb bass response using only cheap 6 watt OPT's. Incredible imaging with no coupling caps or cathode by-pass used at all.
I should probably do some more measurements and maybe create a thread...
Ian
Last edited:
In LW circuit from 1929 capacitors in PSU was 1uF. Then there was no greater value capacitors. That's why he had to invent the PSU noise cancel.
Today it is not necessary.
I would certainly not say "Today it is not necessary". I would rather say "Today it is done in a simple but inferior manner with cathode bypass." 😉
Take a good look at the original schematic. There is also no cathode by-pass on the output either. This makes a huge difference.
I personally find a Cgh (in the original schematic) of over 2uF yields significantly better results than 1uF. So the original design from 1929 was limited more by the materials at hand than at first glance.
Last edited:
Today, tubes aren't necessary.
The LF approach is quite elegant and relevant. True, it's not the easiest or most flexible approach. The input/driver and output stages cannot be independently designed, which limits the degrees of freedom to chose tubes and operating points. Darius's bootstrapped follower addition mitigates this issue some, but not entirely. But once a design is worked out, frequency and phase response is stellar - limited only by the output transformer.
Sheldon
RajkoM - Have you really been on this forum since 2009? unbelieveable! 😉
That is the same year that Darius made his last postings here as OldEurope.. maybe you missed it all.
True that you can't independently design the stages. This is probably one of the reasons few have tried it. The calculations were not hard though (with credit to Sheldon here for kind advice).
The bootsrapped follower variation by Darius works very, very well. I used 6SL7 since I had a few russian ones lying around here.
If you don't need quite that much gain then you can chose the 1929 W-L schematic.
BTW - my new OPT's arrived. They are much bigger than the 6 Watt ones... I need to do some new metal work .... 🙁
Last edited:
I don't see any reason to over-complicate things.
For example, power toroid I use have relatively high stray capacitance. And what?
Line and CLC filters reduce noise and hum anyway down to inaudible level on 96 - 98 W/dB sensitive speakers.
Hello Ciro Marzio !!
You and Cristiano Jelasi wrote a most BEAUTIFUL article, published in 1994 in Sound Practices magazine, on two stage DC amps !! A belated many many thanks !!! You two enthusiasts introduced others, ( myself here in the USA ), to many worthwhile ideas.
I adore two stage DC amps!!
I agree with you about the power supply iron 100%. Here is what I have evolved, over the last 33 years :
(1) I prefer to use custom power transformers, large, with secondary high voltage DCRs of only 9 Ohms, from either end to center tap, ( 18 Ohms across entire winding ).
(2) My B+ filters to the Finals are always L1/C1/L2/C2 with all Ls being small, under 10 Ohms DCR, and also, under two pounds in weight.
(3) No Cs over 50 uF . All Cs in SE amp are film ( the new WIMA DC LINKS are excellent ). All Cs in the SE amp are also multiple bypassed, using MORE than the two 0.22 and 0.01 uF values that you used in 1994.
(4) Large quality wire to hook up power supply grounds and B+, either 13 AWG Siltech ( silver with a little gold, very expensive ) or 12 AWG military wire ( stranded copper, silver plated ).
-----------------------------------------------
Two comments from me ( IMHO ) about the Shishido amp's audio stages.
First, I would not do a direct copy of that audio circuit. The sensitive 12AX7 stage should not use paralleled sections, without a smearing and audible loss in fidelity, compared to just one section as the driver. Secondly, I would reduce the dissipation of the 2A3 plate that Nobu Shishido used, to operate the 2A3 more conservatively...sounds better when all else is attended to.
Very nice to meet you Ciro on line !! 'Have thought of you for decades. Is there any way we can correspond privately by email ? Mine is drlowmu ATA gee mail dottie commie. You can figure that out I suspect.
Jeff Medwin
I built my first L-W back in 1990 while writing my PHD dissertation. Building that old circuit was my way to relax. First (monophonic) prototype was breadboarded in an old drawer and used 6SL7+2A3. Very close to original circuit. No cathode by-pass cap for the 2A3.
I remember that without a cathode bypass cap for the 2A3 the sound was more "compressed/constricted". Adding the cathode bypass everything was more relaxed and natural, even if narrowed the bandwidth a little. But the quality of that cap was VERY critical. I was lucky enough to have a huge CSI 15 microF at hand and I first tried that.
Later I tried something more "normal" in order to keep size smaller but... the magic faded away.
IMHO what I considered - and still consider - the sonic "signature" of L-W, and the main difference between L-W and a "normal" DC coupled SE amp, is the presence of the cap between power supply and output tube cathode (C2 in the original schematic - C4 in SP amp).
That cap changes everything. Electrically and sonically.
Of course the SP L-W could work without C4 but the sounds "collapse", L-W magic disappears and it sounds like any other honest DC coupled SE amp.
Now, after so many years I am thinking of making a L-W amp for my wife (she got a work in another town). I have to keep this small, but I am pretty sure this will use good ole 45 as an output tube.
Thinking of either 6SL7 or a pentode (E180F or EF84 probably) in the input stage. It's time to research L-W deeper 😉
Thanks Jeff, very nice to meet you too!
I guess you can use my profile to contact me by email.
But for good measure I have sent you a PM with mine 😉
I guess I am an annoying perfectionist 😛
You know, being a luthier can be both a blessing and a curse when it comes to audio reproduction.
If something sounds unnatural I can hardly stand it.
IMHO, hum is just a part of the picture.
Low stray capacitance / low flux density TAs allow a more detailed, relaxed, natural sound.
If your first prototype used 6SL7 then I have no doubt that it sounded poor without cathode bypass on the input/driver stage. The problem of using high gain triodes (with high Z) could only be solved that way back then...
You might have noticed that the schematic I use is different than the one you breadboarded in 1990... It solves the problem of High Z without using any cathode by-pass. Seems like only a handful of people have tried it since it was first proposed in 2007 though. That is too bad, since it gets us away from the hocus-pocus one often hears about cathode bypass. Of course, some gurus want to sell little boutique film capacitors for exhorbant sums - that's their business. But you really don't need to buy into all of that. 😉
For C2 I use a 35uF ASC x386 which cost about $12. I agree that its important, which is why I splurged for the ASC and made space for it. 😀
Last edited:
Hello,
You state :
" Seems like only a handful of people have tried it since it was first proposed in 2007 though."
Several QUESTIONS :
I have been THINKING about this !!! OK, "I" would like to learn HOW to implement the circuitry you discuss, Soulmerchant, that you and others use to eliminate the Rk capacitor bypasses.
QUESTION (1) :
Where can I systematically go on line to LEARN how to do this, easily, OR, is there ANY way you, or someone else can personally can show me the way to do the math??
Who can point me in the direction, posts on line, that will address this topology implementation ?
I FULLY understand how to compute a simple direct couple. Doing it without a Rk bypass is my new challenge, hence my sincere request for guidance . I will TRY this.
QUESTION (2) :
Who has compared the same tube line - up in two stage DC amps both ways?
What are the sonic differences when both topologies are each OPTIMALLY implemented?
I am assuming, in the case of film cap bypassed Rks, one spared NO expense on the Rk caps, and had used multiple value caps of highest quality. Doing so, allows one to linearize the amps response on playback.
( It takes, IMHO, a discriminating ear and a large pocketbook to do this, as caps are probably the worst component in audio - and only the best caps seem to be adequate. )
-----------------------------------------------------------
My general tube-choice tendency is to use a high gain driver stage tube like one section ( half ) of either a 12AX7 or a 6SL7, to drive a 2A3 , and sometimes I use a Type 45 as the Final tube.
Thanks in advance, "older dog - seeks to learn new ways" !!
Jeff Medwin
- Status
- Not open for further replies.